A rapid, noninvasive, and accurate, point-of-care device to diagnose and exclude pulmonary embolism (PE) remains an unmet need, especially for patients with conditions that activate inflammatory and coagulation pathways and therefore predispose to PE. These conditions include surgery, cancer, age >70 years, prolonged hospitalization, hemodialysis, connective tissue and infectious diseases. In patients with these conditions, the quantitative blood D-dimer assay consistently shows a diagnostic specificity below 25% for PE, thus limiting its usefulness as a rule-out test in the patients who engender the most worry for PE. Although many diseases elevate the D-dimer, few diseases besides PE obstruct blood flow to large volumes of alveoli. The novel diagnostic device, the Carboximeter(r), measures the effect of high-ventilation, low-perfusion mismatch from PE on the composition of exhaled breaths. The central hypothesis states that the instantaneous ratio of the expired partial pressure of CO2/O2 at the end of the alveolar phase of steady- state tidal breathing (Carboximetry(r)) will accurately discriminate patients with PE from patients who have other conditions that mimic PE. Three prior clinical trials of 370 patients (n=77 PE+) have demonstrated that the CO2/O2 ratio >0.45 has diagnostic sensitivity and specificity at least equivalent to the quantitative D-dimer <500 ng/mL. In symptomatic post-surgical patients, we hypothesize that Carboximetry will be especially useful to monitor for PE if the CO2/O2 ratio could be compared with preoperative measurements: a minimal change in this ratio would suggest absence of PE whereas a significant decrease would suggest PE. Therefore, in phase I, we will measure the pre- and post-operative CO2/O2 ratio and quantitative D-dimer in 100 high-risk orthopedic and oncological surgery patients. Patients will be followed prospectively to establish outcome with respect to diagnosis of PE. The specific aim is to measure variability caused by anesthesia-surgery on the ratio. If the coefficient of variability (CV) for the CO2/O2 ratio is substantially less than the CV for the D-dimer in patients without PE, this will provide initial evidence that Carboximetry can be used as a monitoring device for PE in high-risk surgical patients. In phase II, we will directly test the diagnostic accuracy of the CO2/O2 ratio versus the quantitative D-dimer in 350 symptomatic patients with conditions known to predispose to PE and elevate the D-dimer. All patients in phase II will undergo CT angiography-venography and 30 day follow-up. Sample size is computed to test the hypothesis of diagnostic equivalence between Carboximetry and the quantitative D-dimer for the area under the receiver operating characteristic (ROC) curve. Secondary analysis of phase II will test if the combination of either a CO2/O2 ratio >0.45 or a D-dimer <500 ng rules out PE with sensitivity >98% and specificity >50%. These highly feasibly specific aims will provide background data needed to establish Carboximetry as a viable alternative or adjunct to the D-dimer to screen for PE in high-risk patients. The condition of clots in the lung, termed pulmonary embolism (PE), is the second leading cause of sudden, unexpected death in the U.S. Physicians lack a quick, accurate method to diagnose and exclude PE at the bedside, especially in patients at risk for PE. This project will fund a clinical study designed to test the diagnostic accuracy of a hand- held breath device that detects the altered gas exchange physiology caused by PE in post-surgical and other patients at high risk for PE.